Various data recording, playback, and erasing techniques exist for recording data to and from data storage media, such as magnetic tape. Magnetic tapes are used for data storage in computer systems requiring data removability, low-cost data storage, high data-rate capability, and high volumetric efficiency and reusability. The rapidly accelerating growth in stored digital data and images, the Internet, and replacement of paper as long-term record retention, and the need for massive dense storage for reconnaissance and surveillance is creating a demand for corresponding increases in the data storage capacities of magnetic tape recording and reproducing systems, while maintaining the special requirements of high speed digital tape systems.
Tape recording and reproducing systems for use as computer data storage devices are often required to provide high data transfer rates and to perform a read check on all written data. To satisfy these requirements, conventional, orthogonal linear tape systems (where recorded transition lines that are created between regions of opposite magnetization are orthogonal to the head/tape motion direction) typically employ methods wherein the tracks of data lie parallel to each other and to the edge of the tape. Linear recording techniques offer high data transfer rates by employing reading and writing head configurations with multiple, parallel channels, wherein each read and write head pair provides a channel, typically with each writing or reading element in data transfer contact with the recording media a substantial portion of the time.
In orthogonal linear tape recording systems, data tracks generally are followed in the direction of tape movement with the read and write heads arranged in the same manner as the recorded transitions that are perpendicular to the direction of tape motion. The writer element to a significant degree defines the width of a data track (and thus the number of data tracks that can be provided across a tape of given width) by creating the regions or domains of magnetization following one another in the tape direction at the width of the write head.
The potential for misregistration of the read element to the written track (from tape wander, data track alignment or the like) requires in some systems that the read element be substantially smaller than the written track width in order to ensure that the read head is reading magnetization fields only within the desired data track. Thus, the read head element also (as is also limited by read head performance characteristics) limits how narrow the data track can be, and hence the maximum track density.
Not only is the data track width limited by the minimum read element size in order to meet the recording system's performance criterion, it also is limited to accommodate expected misregistration as may occur under the dynamic conditions of moving media and as may be determined empirically or by modeling. If a read head moves off the data track for whatever reason and begins to read a signal from the adjacent track, the possibility of erroneous data transfer increases. More specifically, the error rate is known to increase exponentially as the read head moves further off the data track. Typically, for an acceptable off-track error rate, the adjacent track signal must be less than ten percent of the data track signal being read.
The general premise is thus to write wide and read narrow. Writing wide, however, decreases the data density (less data tracks across a given tape width). Reading narrow is unfortunately limited by making an acceptable read element that will still meet signal amplitude, SNR (signal to noise ratio), and media defect sensitivity requirements. As a result, minimum track width is approximately the width of a read element that meets the above performance requirements plus twice the misregistration (normally the three sigma value since the misregistration is a statistical distribution).
There are a number of potential sources of read element to written track misregistration error, which error is systematic in that both the media and the drive are involved as potential sources of error. The principal sources of error include tape lateral motion, vibration in the head/actuator assembly, dimensional instability of the media substrate, and mechanical misalignments between read and write elements in manufacturing and assembly. Probably the most significant limitation on tape track densities is the tendency for the tape to experience lateral tape motion, which is a tendency for the tape to shift laterally relative to the linear direction of tape motion. During a data track write, lateral tape motion can cause one or more data tracks to deviate from a desired axis along which tracks are expected to be written. During reading, lateral tape motion can cause misregistration of the read head over the track being read. This build-up of potential misregistration of data tracks combined with other less significant potential sources of misregistration can result in a portion of the read element to be positioned over an adjacent data track, which, if significant enough, can cause an unacceptable level of data transfer errors. As noted above, the reading of an adjacent track is typically limited to ten percent or less of the desired data track signal. The normal method in linear tape recording to ameliorate the potential effects of this misregistration is to make the read element much narrower than, i.e., approximately half, the track width. However, as noted above, limitations of minimum signal amplitude, signal-to-noise-ratio, and sensitivity to media defects provides a lower limit as to how narrow the read element can actually be. Thus, from a practical design perspective, an effective read head size as determined by such performance constraints would be doubled to determine a desirable data track width. As such, the effective read element size limits how narrow a data track can be made.
One developed method of increasing data track density involves azimuth recording techniques. Azimuth recording for data tracks has long been used in helical recording systems and has been more recently introduced into linear tape systems. Generally, in azimuth recording of either helical or linear tape systems, data transitions on alternate adjacent tracks are recorded at a similar but opposite azimuth angle (e.g., θ on one track and −θ on an adjacent track, with this alternating azimuth pattern repeating across the data band) and relative to an axis along which the head travels relative to the media. In helical tape recording systems, the head is moved relative to a linear tape movement at a significantly greater speed and at an angle to the relative direction of tape movement.
Azimuth recording itself is a well-understood technology that provides a level of suppression of an adjacent track signal. The suppression is based upon the well known relationship that the suppression, S=20*log 10 [sin x/x], where x=(πW/λ))*tan 2θ. In this relationship W is the data track width, θ is the positive value of the +/−θ angles that the recorded transitions make with the transverse axis to the head direction, and (is the wavelength associated with the minimum transition density (λ=two times the maximum transition spacing). Thus, a determined azimuth angle, θ, is dependent on factors such as the degree of suppression to be attained, the data track width W, and the minimum transition density or maximum λ of the readback signal spectra. In current systems the data track width W is at least an order of magnitude larger than λ and thus, a suitable transition angle θ can be relatively small to achieve sufficient suppression of an adjacent data track signal.
Because of such angular azimuth recording, a signal from a track adjacent to the data track being read can be sufficiently suppressed to an acceptable level, such as to be less than ten percent of the data track signal as noted above. Hence, a read element can overextend an adjacent track and thus can be designed to be wider than the data track, allowing the full data track signal to be utilized. Azimuth style recording for data tracks is further described in Assignee's co-pending U.S. Pat. No. 6,947,247, the entirety of which is incorporated herein by reference; as well as in U.S. Pat. No. 4,539,615.
Some current linear serpentine tape drives for azimuth recording typically utilize a single head structure that contains two pairs of read and write elements. Like orthogonal head structures, azimuthal head structures are typically designed with the read and write elements parallel to each other and aligned in the direction of tape movement when brought into the proper alignment with the desired azimuth angle. Thus, by offsetting the read and write elements as they are positioned along lines that are parallel to one another as to the distance along the parallel lines, an orthogonally constructed head can be positioned to record and read azimuthal tracks when rotated at an appropriate angle. The read and write elements can be aligned so that with the proper spatial relationship between them, they are able to read and write adjacent tracks and only require transversal repositioning once for every track pair. Such transversal movement and positioning or tracking can be conventionally controlled by known actuators. Tracking can be achieved in a single head, but usually requires the additional complexity and weight of a dual degree freedom actuator, such as conventionally known and that permits both rotary movement of the single head and movement of the head in the transverse direction to the tape movement. A compound dual degree freedom of motion actuator, i.e. a single unit to provide multiple types of motion, adds additional mass and generally needs to carry twice as many leads in order to accommodate forward and reverse read and write capabilities. This provision of additional leads adds stiffness to the system that can inhibit or interfere with its motion capabilities.
Recent generations of multi-channel linear serpentine tape systems have used servo tracking to decrease track misregistration. The use of servo tracking has greatly reduced tracking errors due to manufacturing alignment and offset tolerances between the read and write element arrays, skew errors, some track shift due to tape substrate dimensional instability, and the effect of lateral tape motion. In such systems, position sensing read sensors (servo elements) detect prewritten servo tracks on the tape that can be laid down under tightly controlled conditions to reduce misalignment of the servo tracks to the tape. The tape is typically divided into alternating bands of data tracks and servo tracks where the band of data tracks can be much wider than the servo band; typically the data band is 8 to 16 times the width of the servo band, depending on the number of data channels. From the output signals of the servo data elements, a position error signal can be determined that is used by the servo control loop to dynamically and more accurately position the data elements over their tracks. Typically, the servo elements are located in the same array as the read elements and can be symmetrically placed outboard of the read array on each side.
Notwithstanding the widespread use of servo systems and formats, in helical recording the Position Error Signal (PES) generally has been embedded in the data-recording band and uses the data read head as the servo transducer. Also, when recording or writing, the head moves in only one direction relative to the tape and the tape is only moving in one direction. Quantum Corp., for example, has used azimuthal recording in its DLT drives, but does not track follow.
A number of different encoding schemes have been proposed for servo formats. The four most prevalent forms of encoding are frequency encoding, amplitude encoding, time-base encoding, and phase encoding. All tend to share a common characteristic where the servo transducer is a single element. Further, except for time-based encoding, the primary characteristic of these approaches is that the encoded servo features on alternate servo tracks are different. In some cases this differentiation can be extended to more tracks to provide either a larger capture range when accessing the track or enhancing track identification.
Most current servo systems used commercially in linear serpentine tape systems commonly employ either an amplitude modulated mono-frequency signal (AM system) or a “Time Base” system. A typical AM system might utilize a single servo read element to detect the position error signal where the “on track” PES is half, or less, than the data signal. The weaknesses of the AM approach include the susceptibility of the PES to dropouts and noise, the reduced sample rate, the wider band width to accommodate the modulation, and without writing (erasing the holes) the tracks individually, the inability to identify the selected track.
Time-based servo position error signals have been introduced by IBM in some of its latest products and the philosophy has been extended to the LTO family drives that are being produced by IBM, Seagate, and HP. Time-based servos use slightly (typically 6-7 degrees) but, oppositely angled transitions for the servo timing features, e.g., “diamond-shaped,” “vee,” “inverted vee” features, combinations of these, or the like. The time difference among servo transitions as a function of transverse position of the servo head on tape provides the positioning information. The servo transducer orientation is nominally perpendicular to the track direction so that the transitions are encountered at a slight angle.
Hybrid thin film/ferrite servowriter heads with precision patterns have been developed to record the time-based servo tracks for IBM and LTO tape heads. See, e.g., U.S. Pat. No. 6,795,246, incorporated herein by reference in its entirety. See, also, U.S. Pat. Nos. 6,496,328; 6,269,533; 6,678,116; 6,894,869; and 6,989,960, all of which are incorporated by reference herein in their entireties.
A typical Time Base system might use a servo read element much narrower than the track width or data read elements, hence a much lower signal level and signal-to-noise ratio (SNR), while trying to achieve a high spatial resolution. Neither system typically provides positive track identification, although the Time Based system could allow servo group identification.
A significant advantage of the “Time Based” Servo approach is that it is relatively insensitive to dropouts and Gaussian noise. However, because of the narrowness of the servo transducer, the signal to noise ratio (SNR) is quite low. Perturbation along the axis of the tape in the transitions is magnified in the transverse direction by 1/sin (, where (is the azimuth angle of the transitions. Thus, with an azimuth angle of about 6 degrees to 7 degrees, perturbation is magnified by a factor of about ten. Further, although the measurement is insensitive to any static variation in the tape speed, significant error may tend to be introduced by any dynamic variation in the tape speed. Like the AM system, this approach typically does not provide for unique identification of the data track. Further, modeling and simulation have shown that the current time-based approaches may be limited to intrinsic (to the servo pattern and head only) misregistrations of several tenths of micron, thus limiting maximum track density to 4,000-6,000 tracks per inch.
With the trend toward recording higher densities, the industry strongly needs a servo format and system that allows increasingly more accurate on track guidance. It would be further desirable to have a servo format and system that allows positive track and group identification at the beginning, end, and optionally periodically along the length of a tape.